The Real Estate Risk Management Process:
Integrating Tools from Other Disciplines
by
Paul E. Adornato
Bachelor of Arts
Dartmouth College
(1986)
Submitted to the Department of Urban Studies and Planning
in Partial Fulfillment of the Requirements for the Degree of
Master of Science in Real Estate Development
at the
Massachusetts Institute of Technology
July 1992
(Paul
E. Adornato 1992
All Rights Reserved
The author hereby grants to MIT permission to reproduce and to
distribute publicly copies of this thesis document in whole or in
part.
Signature of Author
Department of Urban Studies and Planning
July 31, 1992
Certified by
Thiomas A. Steele III
Chairman, Center for Real Estate
Accepted by
Lawrence S. Bacow
Chairman, Interdepartmental Degree Program in
Real Estate Development
Rotch
MASSACHUSETTS INSTITUTE
OF TEHNO, O(Gy
SEP 02 1992
LIBRARIES
The Real Estate Risk Management Process:
Integrating Tools from Other Disciplines
by
Paul E. Adornato
Submitted to the Department of Urban Studies and Planning in
Partial Fulfillment of the Requirements for the Degree of
Master of Science in Real Estate Development.
ABSTRACT
Risk, defined as the volatility of investment returns, is an
important consideration in the real estate decision-making
process. Current real estate risk management models are
largely based upon securities investment models of Modern
Portfolio Theory (MPT). However, these models have two
principal drawbacks: they ignore some of real estate's
unique characteristics, and they require data which are
unavailable or difficult to estimate for real estate. This
thesis introduces tools from other disciplines to model the
early stages of the risk management process: risk
identification and categorization.
Four distinct models are explored. Chapter 2 applies the
theory of comparative advantage to real estate risk and
suggests an analytical tool for risk management. Chapter 3
considers the options characteristics of real estate and
explores how options pricing models may apply to real estate
valuation. Chapter 4 notes empirical findings which
question the applicability of expected utility theory and
highlights the advantages of explicitly considering downside
risk. Chapter 5 reports on a generalization of a widelyaccepted investment model, the Capital Asset Pricing Model
(CAPM), which may be more suitable to real estate than its
more narrow form. Chapter 6 concludes.
Thesis Supervisor:
Thomas A. Steele III
Chairman, Center for Real Estate
Table of Contents
2
...........................
..........................
Abstract
Chapter 1: Real Estate Risk............
Chapter 2: Comparative Advantage in Real Estate ........... 14
Chapter 3: The options Characteristics of Real Estate.....25
33
Chapter 4: Modified Expected Utility Theory.. .............
Chapter 5: A Consumption-Based CAPM ....................... 44
Chapter 6:
Bibliography
Conclusionsn...............
..............
..............
.....................
51
..........---.----
54
Exhibits
1.1 systematic and Non-Systematic Risk...........
1.2 Sources of Return Variation..................
1.3 The Risk Management Process..................
2.1
. ..
. ..
. ...
..
4.2
Plot of Real Estate Participants............. .........
Plot of Components of Real Estate Risk....... .........
Classic Expected Utility Theory..............
Probability Distribution of Expected Returns.
4.3
Portfolio
4.4
Volatility Around the Mean...................
Comparison of Risk Measures..................
Capital Asset Pricing Model..................
2.2
4.1
4.5
5.1
...
5
. .. 7
7
17
19
Returns ............................
.42
46
Chapter 1
Real Estate Risk
Chapter Summary: The concept of risk in real
estate investment and development is defined.
The risk mitigation process is outlined. The
shortcomings of current real estate risk
mitigation strategies are noted, and a
framework for different risk identification
and categorization techniques is set.
1.1 Definition of Risk Risk is common term in everyday
language which has a variety of connotations.
Risk can be
defined in a number of ways, such as: the probability of
loss, the probability of not receiving what is expected, or
the probability that an investor will not achieve his
required rate of return on an investment.
Real estate investment and development is a complex
activity which usually spans many months or years.
Consequently, real estate investment returns tend to be
volatile; that is, returns often vary from period to period.
This variation is often difficult to predict.
For the
purposes of many real estate professionals (and for the
purpose of this thesis), risk is defined as the possibility
that actual returns are different from expected returns.
This difference may be positive or negative.
Real estate
returns vary from period to period due to a variety of
causes, and it is the goal of this thesis is to introduce
tools to better define and evaluate real estate risk.
Just as there are many ways to define risk, there are
numerous ways to categorize elements of risk.
Many real
estate professionals first categorize elements of risk into
systematic risks and specific risks.
Systematic risks are
those risks which affect the entire system or universe of
which the investment is a part.
Since these risks are
exogenous to the system and affect all members of a
particular universe, it is not possible to avoid these
risks.
An example of a systematic risk might be inflation,
since inflation would affect all real estate investments,
and furthermore, inflation is considered beyond the control
of individual investors.
Exhibit 1.1: Systematic and Non-Systematic Risk
Real Estate
Return Risk
Systematic Risk
Non-Systematic Risk
*
Exogenous
* Endogenous
*
Non-Diversifiable
* Diversifiable
Alternatively, leasing risk is an example of a specific or
endogenous risk.
Leasing a project is generally within the
control of the developer of the project who can use his
skill and knowledge of the market and of potential tenants
to secure the highest rental rates, the longest lease terms,
or the most creditworthy tenants possible.
The division between exogenous and endogenous risks (or,
non-diversifiable and diversifiable risks) is often blurred
and depends on the boundaries of the universe of the case.
For example, a person who invests only in real estate may
consider real estate market values an exogenous risk, since
variability in the level of real estate values is beyond his
control.
Alternatively, another investor who owns a
portfolio of investments including stocks, bonds, and real
estate may not be as impacted by volatile investment returns
because she owns assets other than real estate.
Real estate
market variability is endogenous to her because her
portfolio returns may not vary due to changing real estate
market values; other assets' returns may offset real estate
market variability.
Thus, the same element of risk, real
estate market variability, may be exogenous to one investor
and endogenous to another.
1.2 Types and Sources of Risk
There are three components of
real estate investment returns: the capital investment, the
operating cash flows, and the residual asset value.
component can be volatile.
Each
Capital investment variations
may affect the construction or renovation expenses and alter
investment return.
Cash flow variations are changes in the
ongoing operating return stream.
Residual value variations
arise from appreciation or depreciation of the market value
of the property.
Exhibit 1.2: Sources of Return Variation
Operating
Cost Risk
Financial
Leverage
Market
Effects
Capital
Market
Effects
Regulatory
Effects
Real Estate Return Variation]
These variations of real estate investment returns have
many sources.
These are some major categories of sources of
variation:
Operating cost risks include those sources of variation
in the cost of providing the day-to-day services to
operate a property, such as janitorial services and
security services.
Financial leverage exposes investment returns to
movements in interest rates or changes in financing
terms. The amount of debt and financing terms
determine the direction and magnitude of return
variations.
Market effects such as the local supply and demand for
space affect real estate returns.
Capital market effects such as the terms and
availability of investment monies from banks, insurance
companies, pension funds, and the public securities
market are likely to affect real estate investment
returns.
Regulatory effects such as investment tax credits for
low income housing can also greatly influence
investment returns.
1.3 The Risk Management Process
Given the complexity of
real estate investment, it is important to approach risk
management in a consistent and logical manner.
Just as
there are different ways to define risk and to describe the
sources of risk, there are various methods to manage risk.
It is important to note that risk management is an ongoing
process which should be reviewed frequently, as the sources
of risk continually change and interact.
The steps in the
risk management process are as follows:
Identify the broad concepts of real estate investment
risk, such as systematic risk and specific risk.
Categorize the risks by their source.
Analyze the risks in each category to determine their
characteristics, magnitude, and sensitivity under
different scenarios.
Mitigate the risks by carrying out specific strategies,
such as diversification under modern portfolio theory
or structuring the deal to hedge or to shift risk.
Realize investment returns. The process should be
continually reviewed so that risk mitigation strategies
can be adjusted to reflect current conditions.
Exhibit 1.3: The Risk Management Process
Identify
Categorize
Mitigate
Analyze
1.4 Risk Mitigation and Modern Portfolio Theory
Realize
Modern
Portfolio Theory (MPT) is a collection of investment
principles which has gained wide acceptance by institutional
investors over the last 20 years.
MPT is considered the
standard investment philosophy among institutional stock and
bond investors.
The first principle of MPT was described by
Harry Markowitz'
in
1959; it
"efficient frontier."
the concept of the
is
Markowitz held that in order to
minimize variability of investment returns between periods,
investors should own a portfolio of investments whose
returns are not correlated. Further, the efficient frontier
represents a trade-off between risk (variability of returns)
1
, Portfolio Selection - Efficient Diversification of
Investments, 1959, Yale University Press, New Haven.
9
and expected rate of return which maximizes the expected
return for each level of risk.
Theorists have extended MPT
with other investment models.
Applying MPT to stocks and bonds was relatively
straightforward, given the liquidity of the securities, the
frequency of transactions, the substantial historical
trading data, and the overall efficiency of the securities
markets.
Soon institutional investors extended the
diversification theory to argue in favor of real estate
investment. Real estate was included in the investment
portfolios of institutional investors in order to further
diversify their portfolio returns beyond stocks and bonds,
due to preliminary data showing low or negative real estate
return correlation to stocks and bonds.
Once large-scale institutional real estate investment was
accepted, academicians and practitioners sought to apply MPT
within their real estate portfolios.
Thus, the focus of
applying MPT to real estate was narrowed; that is, the
diversification principles of MPT were applied within the
real estate portfolio rather than viewing real estate as one
component of a larger portfolio consisting of stocks, bonds,
and other asset classes.
The nature of real estate as an investment vehicle has
presented a number of problems in applying MPT.
The data
required to rigorously implement MPT is lacking for real
estate.
Standardized operating data and reliable capital
appreciation estimates are missing, while the unique
characteristics of real estate investments and the
illiquidity of real estate further complicate investment
analysis for real estate.
2
A recent article by Eugene Fama and Kenneth French casts
serious doubt on the usefulness of the capital asset pricing
model (CAPM) as an investment tool. The CAPM is widely used
by securities analysts to assess the risk of equity
portfolios, and its extension to real estate has been a
major goal of real estate investment professionals.
(See
Chapter 5 for a complete discussion.) The researchers
present empirical evidence for stocks which shows that size
(measured by the dollar value of traded stock) and book-tomarket equity are better predictors of stock variation than
beta (historical variation from market returns.)
By ques-
tioning the usefulness of beta within the CAPM for equity
investment management, the paper may also undermine the
foundation of one of the major ideas in real estate investment management.
2
,
Real estate academics and professionals
"The Cross Section of Expected Stock Returns," J
of Finance, 1992, v.47, n.2, pp.427-465.
11
have devoted considerable attention to devising a suitable
beta, or measure of variability from the market, for real
estate.
1.5 Risk Identification and Categorization Revisited
Despite attempts by real estate academics and practitioners,
there does not exist a widespread, practical model to manage
real estate risk.
This paper attempts to survey the
literature from fields other than real estate investment in
order to introduce notions of risk identification and
categorization from those disciplines to real estate
investment. Structuring less developed country investments,
venture capital investing, option theory, and modified
expected utility theory are some fields which may provide a
helpful framework in which to consider the identification
and categorization of real estate investment risk.
Rather than propose a single model for risk management,
this thesis will survey several separate avenues which may
suggest direction for further research.
Chapters 2 through
5 each adapt a concept of risk from another discipline to
real estate.
These models attempt to stimulate thought in
the early phases of risk management: the risk identification
and categorization stages.
Chapter 2 applies a fundamental
economic principle, the theory of comparative advantage, to
real estate risk management.
Chapter 3 considers financial
options theory in relation to real estate.
Modifications to
expected utility theory and downside variance are explored
in Chapter 4, and a more generalized version of the CAPM is
described in Chapter 5.
observations.
Chapter 6 provides concluding
Chapter 2
Comparative Advantage in Bearing Risk
Chapter Summary: Comparative advantage is a
fundamental economic principle which provides
the basis for specialization and trade
between parties. Real estate development and
investment is a complex activity in which
numerous entities interact. Applied to real
estate, the concept of comparative advantage
in bearing risk provides a useful framework
in which to identify and categorize elements
of risk in a multi-party context.
Furthermore, the framework can provide a
basis for devising strategies to mitigate
those risks.
2.1 Introduction
The concept of comparative advantage was
applied to structuring alternative financing arrangements
for investors in less developed countries by Donald R.
Lessard3 .
The idea is extended to real estate in this
chapter by providing a map of how real estate fits the
concept, providing some examples, discussing how the concept
can be applied within the context of modern portfolio
theory, and showing what the framework implies for deal
structure.
2.2 The Theory of Comparative Advantage
Eighteenth century
economic theorist David Ricardo first described the theory
3
,
"Alternative Finance for Less Developed Countries:
A Primer," 1991, unpublished working paper, MIT.
of comparative advantage, a fundamental economic principle
which has important implications for specialization and
trade.
example.
The theory is easily understood by way of a simple
Suppose that two countries, Japan and China,
produce only two goods, televisions and rice.
Japan can
produce televisions more efficiently than China, and China
can produce rice more efficiently than Japan.
Each country
has an absolute advantage in the production of one good.
It
is to Japan and China's mutual advantage to specialize in
producing the good which they can produce most efficiently
and to trade for the good they lack.
In this way, each
country is better off producing what it produces best, and
trading some of the surplus for their neighbor's product.
Suppose, however, that Japan is more efficient in making
televisions and growing rice; that is, Japan has an absolute
advantage in producing both goods.
China can produce both
goods, too, but it is extremely inefficient in producing TVs
but only slightly less efficient than Japan in growing rice.
Both countries can be better off (that is, the sum of all
goods produced in both countries is maximized) when each
specializes in what it has a comparative advantage.
Japan
should produce only televisions and sell some to China.
China is less inefficient in producing rice than it is in
producing televisions. Japan's gain in producing televisions
more than offsets China's loss of producing rice, and with
trade, both countries are better off.
2.3 Comparative Advantage in Real Estate
Real estate
investment and development is rarely a single-party
enterprise and often involves many important participants.
Various entities supply one or more inputs in the
development, operation, and investment process.
Such inputs
include investment capital, design expertise, construction
knowledge, market research, brokerage services, management
of the development process, management of the operational
phase, and other factors of production.
Each participant
brings a set of skills and resources which may or may not be
deployed in any particular project. There is significant
diversity in the goals, skills, resources, and experiences
of the participants.
To understand the array of potential resources represented
by participants in a development or investment project
requires explicit differentiation. One way to compare the
participants is to plot them in a three dimensional chart,
as shown in Exhibit 2.1.
The axes used to place the
participants are: the ability to diversify, the ability to
influence project outcome, and the level of information
regarding the project.
These axes are now more fully
defined.
16
For an investor, the ability to diversify can be thought
of as the ability to hold assets of different types or
geographic locations, or the ability to hold a diverse
portfolio of stocks, bonds, and other investment vehicles.
For a construction firm, ability to diversify might mean the
ability to engage in projects spanning different time
periods or the ability to seek projects of varying
construction types (industrial site-cast buildings and
multi-family masonry structures.)
Exhibit 2.1: Plot of Real Estate Participants
Ability to Bear Project Risks
Individual Investors
Limited Partnerships
Commercial Banks
Pension Fur
lopers
Information Regarding
Project Prospects
Ability to Influence
Project Outcome
The ability to influence project outcome, the second axis,
is fairly self-explanatory.
It is the ability to influence
any of the three characteristics of investment return from
Chapter 1:
the capital investment, the ongoing cash flow,
or the residual value.
A property manager may have the
ability to keep operating costs low, while a large insurance
company may be able to provide a favorable financing
package.
The third axis, the ability to gain information regarding
the project prospects, may be thought of as the ability to
predict outcomes which might affect the project.
A large
brokerage firm may have information regarding future market
demand, rental rates, and tenant needs.
An architectural
firm may know how a project design can accommodate future
telecommunication or computer expansion lines.
As illustrated in the diagram, participants of a specific
type tend to occupy certain sections of the diversificationinfluence-information space; however, among different players of the same general type, there may be considerable
variation in position. The size of the firm, the capital
resources and the human talent all affect firms'
delineation.
The components of risk, such as operating risk, financial
leverage, local market effects, capital market effects, and
regulatory effects, can be similarly graphed in the three
dimensional space.
As illustrated in Exhibit 2.2, the
position of the sources of risk may vary substantially for
different participants.
Exhibit 2.2: Plot of Sources of Real Estate Risk
Ability to Bear Risk (Diversify)
Financial
Leverage
Operating
Risks
Capital Market Risks
Regulatory
Risks
Market
Risks
Information
Ability to Influence
Outcome
For example, one development company may be a large
nationwide firm with considerable experience and expertise
building large projects throughout the country. It may have
relationships which allow it to gain an informational
advantage or which allow it to exert influence over other
participants.
Alternatively, another developer may be a
smaller regional player which seeks a project outside of its
primary market area where its information and ability to
influence are relatively low.
This would suggest that the
same project would have a different configuration of the
components of risk in the three dimensional space, based on
the individual firm's characteristics and expertise.
In the
matrix, it is better to be far from the origin (the point of
greatest risk.)
In a similar fashion, the traditional components of real
estate risk may be evaluated from the perspective of the
individual parties.
Factors such as inflation, tax effects,
financing, leasing, and market risks may be either
endogenous (specific) or exogenous (systematic) to different
players depending upon their size, experience or expertise.
While the large national developer may consider regional
economic trends diversifiable, a small regional developer
operating entirely within the region may not be able to
diversify away from the effects regional economy.
same potential project, each developer would have a
20
For the
different matrix, and should seek other participants -financial partners, contractors, etc. --
who have the best
complementary "fit" with their project risk matrix.
In current real estate investment literature, the analysis
often begins after the categorization into endogenous or
exogenous risks has been made.
This approach avoids
explicit acknowledgement of the richness of the
possibilities, thereby neglecting the relative nature of
risk categorization.
This, in turn, limits the range
possible strategies for mitigation of risk.
2.4 Real Estate Applications
Applying the concept to real
estate development and investment confirms why deal
structuring plays such an important role in risk management.
Development deals or real estate investments should be
"engineered" so that ideally the party best able to mitigate
a particular risk bears that risk.
For example, in a joint venture between a developer and a
financial partner, the developer should bear the risk of
(and receive a reward for) the maintaining the construction
budget, since the developer has the power to control
subcontractor construction costs and oversee the day-to-day
construction process.
Likewise, the financial partner
presumably has substantial capital resources and cash
management expertise, which implies that it should be able
to diversify and hedge interest rate fluctuation.
Even
inflation, which is typically considered an exogenous risk
to most real estate market participants, may be better
tolerated
by international investors who may diversify
globally.
The crisis unwinding in the real estate and savings and
loan industries provides an example of poor allocation of
risk.
Developers eager to build and bankers eager to lend
found a mutually beneficial partnership in which profits
from development and lending exceeded the perceived risks.
Depositors did not need to scrutinize the lending activity
of their bank, since deposit insurance eliminated their
risk.
The Federal government was an absentee partner in the
deals, and in the form of low cost deposit insurance,
assumed much of the risk of real estate development from all
of the other participants with little ability to control or
mitigate that risk.
2.5 Implications for Deal Structure and Evaluation
The
framework suggests that potential parties in a development
or investment deal not only analyze the deal in the context
of their optimal portfolio and preferred risk profile but
also consider the risk mitigation ability of their coparticipant. Those risks which cannot be borne efficiently
22
by one participant should be shifted to another party. This
"risk engineering" is the key risk mitigation strategy of
the comparative advantage framework. In this way, the sum of
the risk of all parties is minimized.
Under comparative
advantage, the sum of the risk of both parties is minimized
when each party assumes the risk it is best able to manage.
The risk engineering aspect of comparative advantage, in
which participants actively seek to construct a portfolio of
investments or business relationships, implies that the
nature of prospective real estate deal evaluation may
change.
In addition to finding the team of participants
which can best bear the project risks, each participant must
incorporate the project within its portfolio of projects.
The matrix analysis which determines the optimal fit between
parties can also be helpful in evaluating how the project
fits the rest of the portfolio.
2.6 Integration within Modern Portfolio Theory
The
comparative advantage framework is consistent with risk
mitigation according to the principles of modern portfolio
theory.
As currently practiced by institutional real estate
investors, MPT aims to reduce the volatility of the
portfolio return by including investments in the portfolio
whose returns are not correlated.
Risk engineering and the
comparative advantage framework may enhance the portfolio
returns by allowing investment managers to structure a deal
to fit the goals of entire portfolio.
Thus, the concept not
only enhances for portfolio diversification to mitigate
risk, but also suggests risk engineering strategies to
select an individual's optimal risks.
2.7 Conclusion
Comparative advantage in real estate
provides an additional framework through which to consider
risk.
The concept has appeal because it explicitly
evaluates risk from the risk bearer's perspective.
The
categorization provides a starting point to perform risk
engineering to mitigate risk, the next step in the risk
management process.
With each party bearing the risk it is
best able to mitigate, the technique implies that the total
risk of the deal is minimized.
Additionally, the
comparative advantage concept is not inconsistent with MPT
and may enhance attempts to apply MPT by engineering a
portfolio which achieves the diversification goals of MPT.
Chapter 3
The Options Characteristics of Real Estate
Chapter Summary: Real estate investment and
development analysis uses securities
investment models which ignore some unique
characteristics of real estate, namely the
ability to delay or modify decisions. Option
pricing models may be extended to real estate
and can be useful to real estate decision
makers in identifying and categorizing risk.
Traditional Net Present Value (NPV)
3.1 Introduction
analysis holds that a project should be undertaken if the
probability-adjusted present value of all expected cash
flows, discounted at an appropriate rate, is greater than
the project cost.
However, there are other characteristics
of real estate investment that make this decision rule less
than complete.
Two of these characteristics are:
expenditures that are largely irreversible, and decisions
that can often be delayed.
Real estate decision makers often use securities
investment models or capital budgeting models, which do not
emphasize the unique characteristics of real estate. By
considering models which incorporate these unique
characteristics, real estate decision makers may better
understand the elements of risk in the identification and
25
categorization phases of risk management.
Robert S.
Pindyck4 applies option pricing theory to investment
decisions and extends the model to consider a multi-period,
multi-option case.
The extensions capture some unique
characteristics of real estate which traditional securities
investment models ignore.
3.2 The Irreversible Nature of Real Estate
Although most
types of institutional-quality real estate, namely office,
industrial, retail, and multi-family residential, are long
lived assets, the notion of irreversibility has some
relevance to real estate.
Office and retail space often
require significant capital expenditure to refit them if
their is a change in tenancy.
Industrial space is
frequently built to the specifications of a particular
industry or production process, and would lose value if it
had to be adapted to another user.
Repositioning multi-
family residential units by reconfiguring unit layouts or
making other structural changes similarly can incur
significant costs, which are sometimes borne by the
investor.
Although some value remains in repositioned real estate
projects, the expected profit margin may shrink
4
,
"Irreversibility, Uncertainty, and Investment," J.
of Economic Literature, 1991, v.29, pp.1110-1148.
26
significantly or disappear.
Even if improvements to the
land can be reused or adapted for a different use, there may
often be significant erosion in value to the initial
investor.
The case can be made that real estate may lose
much of its value if the original user or intended function
changes.
In this way, real estate development resembles a
sunk cost.
3.3 Options in Real Estate
Many real estate development
decisions clearly have the characteristics of financial
options.
Site control is often obtained by paying the owner
for an option to purchase the land within a specified time
period for a specified price.
This is similar to a call
option in securities markets, in which the holder of the
option retains a right to purchase a specified security at a
specified price within some predetermined time period.
There are other aspects of real estate development and
investment which have characteristics of options.
Consider
that projects can be built in phases in order to obtain more
information about market demand before committing more funds
to the project.
Developers of residential subdivisions can
gain insights into the home buyers' preferences before
building later stages.
Suburban mall developers often own
adjacent land and may choose to expand the mall if
conditions are favorable.
Thus, many development decisions may be delayed to gain
more information and to reduce uncertainty about the
project. In so doing, developers undertake a project when
the risk profile of the project meets their desired risk
profile. Market uncertainty, a systematic risk, becomes more
of a specific or diversifiable risk when more information is
obtained.
Thus, although the risk is still present, there
may be more effective alternatives to mitigate that risk
later in the risk management process.
Similarly, investors may choose to purchase a fully leased
and operating project rather than invest during the
development stage in order to achieve a lower risk profile.
In a real estate portfolio context, the volatility of
returns and the correlation of returns to other assets of a
new project may be unknown, so waiting to observe actual
performance provides insight into the project's pattern of
volatility.
The decision to delay an investment decision
may be considered a risk mitigation strategy.
3.4 A Numerical Example
The concepts introduced thus far
can be made clear through a simple numerical example.
Consider a retail development in which the cost of
construction is $1,000,000, construction is instantaneous
(that is, construction is completed this year, Year 0) and
the rental income realized from the development will be
28
$100,000 in the first year.
Next year, the rental income
will either rise to $150,000 or fall to $50,000 per year,
and remain at that level forever.
The probability of a rise
is 50%, and the probability of a fall is 50%.
assume a 10% discount rate.
Further
Calculating the NPV yields a
result of:
-1,000 + E (
t=o
1 0 0
/(
1
.
1
= $100
)t)
The NPV is positive, so the project should be undertaken.
This conclusion is misleading because it ignores an
option, the option of waiting one year to see if the market
rental rates will increase or decrease.
The cost of this
option is the opportunity cost of not investing now (that
is, not receiving the first year's rental income) and
possibly missing the "window of opportunity" due to the
entry of other competitors.
To see this, consider the NPV
of waiting one year and investing only if the market rental
rates increase.
Note that in Year 0 there is no expenditure
and no income, because we have decided to wait.
In Year 1,
there is only expenditure if the rental income increases.
This has a probability of 50%.
.5
[-1,000/1.1 + E ( 1 5 0 /(
t=1
1
.
1
)t)]
= $295
The NPV today is higher ($295 instead of $100) if we wait
one year before making the investment decision.
clearly better than investing now.
29
Waiting is
Note that if the only choice was between investing today
or not, we would still invest today.
If we had the option
to wait until next year to make the decision, we would
prefer to wait.
to invest later?
What is the value of having the flexibility
The option must have some value because
there is a higher NPV when waiting than not waiting.
The
value of the flexibility is simply the difference in the
NPVs, namely, $295 - $100 = $195.
The cost of the option is
the opportunity cost of not completing the project in the
earlier period: foregone income, possible loss of favorable
financing terms, or discouraging the entry of competitors.
Pindyck also proves that traditional securities option
pricing models give the same result as the NPV analysis.
Although these option pricing models hold appeal as a
framework in which to evaluate real estate risk, there are
some conceptual omissions in applying them to real estate.
Option pricing real estate assumes that a portfolio
consisting of the hard asset (the retail center) and the
income stream (from the rentable square feet) can be
rebalanced from period to period.
In real life it is
difficult to imagine easily changing a project's rentable
square feet.
Majd and Pindyck 5 introduce three characteristics to the
options pricing model which makes it much more appealing for
real estate applications.
First, in the modified model,
both investment decisions and cash outlays occur
sequentially over many time periods.
The decision to
continue investment takes place in stages as more
information becomes available.
properties of a complex option.
This gives the project the
Second, there is a maximum
rate at which cash outlays and construction can occur; that
is, the project takes time to build.
Third, the project
does not yield any cash return until it is completed.
The investment decision model has two input variables: the
total amount of investment remaining for completion, and the
current market value of the completed project.
The control
variable is the rate of investment. The problem is to solve
for the rate of investment which maximizes the value of the
completed project.
Majd and Pindyck solve the equation and
demonstrate how the total value of the development program
can be determined for various remaining amounts of
investment and various market values of the completed
The work is important because it represents a
project.
realistic model to quantify real estate investment risk.
s
,
"Time to Build, Option Value, and Investment
Decisions," J. of Financial Economics, 1987, v.18, pp.7-27.
31
3.6 Conclusion
The modified options decision model holds
appeal as a tool to evaluate real estate risk identification
and categorization.
Investors can consider elements of
risk, such as future market rental rates, construction
costs, or interest rates, in the context of delaying
investment decisions.
The model holds practical appeal as well, because the
estimates required to produce a sensitivity analysis are
typically made in traditional NPV analysis.
The model also
holds promise because it explicitly considers elements of
real estate development and investment, namely the multiperiod construction phase and the multi-option decision
array.
The model departs from the current focus of real
estate investment theory by shifting estimates away from
real estate market estimates (estimating a real estate beta)
to property-based estimates (estimating a range of potential
project costs and returns.)
Chapter 4
Modified Expected Utility Theory
Chapter Summary: There is empirical evidence
which suggests that classic expected utility
theory does not accurately model investment
behavior in some extreme situations.
Modifications of classic expected utility
theory are proposed and described, and their
implications for real estate risk
identification and categorization are
discussed.
4.1 Introduction
Classic expected utility theory has long
been a part of investment decision making, including real
estate investment decisions.
18th century theoretician
Bernoulli described the way people made decisions under risk
or uncertainty by quantifying possible investment returns
and assigning probabilities to those outcomes.
The product
of the possible outcome and the probability of occurrence
yields the expected utility, and the decision maker would
choose the project which maximizes the expected utility.
Exhibit 4.1 provides a simple illustration.
Exhibit 4.1: Classic Expected Utility Theory
Project A
Project B
50% chance to earn $1,000
50% chance to earn $0
60% chance to earn $900
40% chance to earn $0
Expected Utility
Project A
Project B
(.5)x(1,000)
(.6)x(900)
+
+
(.5)x(0)
(.4)x(0)
$500
$540
=
=
Project B has a higher expected utility, therefore, all
other things equal, it is preferred over Project A.
4.2 Modifications
Recent empirical research by Kahneman and
Tversky 6 suggests that expected utility theory does not
adequately model some special cases of decision-making.
They have described several effects which suggest that
classic expected utility theory may not hold in boundary
situations.
The first anomaly is called the "certainty effect."
Subjects were asked to choose between (a) an 80% chance of
winning $4,000 or (b) a 100% chance of winning $3,000.
More
than 80% of the experimental subjects chose option (b) even
though the expected utility of (b),
(1.0) x (3,000)
=
$3,000, is less than the expected utility of option (a),
"Prospect Theory: An Analysis of Decision Under
Risk," Econometrica, 1979, v.47, pp.263-291.
6
,
34
(0.8) x (4,000) = $3,200.
Thus, given the choice between a
greater but uncertain gain and a lesser but certain gain,
most subjects in the sample tended to be biased toward riskaversion, weighting the certainty of gain more than the
magnitude of potential gain.
The certainty effect suggests that there may be systematic
mispricing of assets by market participants. This means that
"risky" investments (those with volatile returns) may be
priced cheaply; that is, an investor who purchases such an
asset may receive returns which more than compensate for the
higher level of risk or volatility because other investors
avoid them and drive the price down.
For example, if
institutional real estate investors shun hotel investments
because of the high volatility of returns, those investors
who do purchase hotels will receive returns which
overcompensates them for the amount of risk they have
assumed.
A second interesting anomaly is called the "reflection
effect," which deals with risk preference in an environment
of likely loss.
Faced with the choice between (a) a loss of
$4,000 with 80% probability or (b) a loss of $3,000 with
100% probability, most test subjects chose (a) even though
its expected loss ($3,200) is greater than option (b)'s
35
$3,000 loss.
Apparently, most subjects became more risk
seeking when the prospect of loss was high.
Again, this finding may have relevance for real estate
investors.
Faced with the prospect of realizing returns
slightly below expectations or making an additional capital
investment (with potentially larger loss) to possibly
achieve expected returns, the theory suggests that
developers would choose the latter, all else equal.
These
two modifications suggest that there may be a bias to behave
differently from what the decision rules prescribe.
In terms of assembling a portfolio of investments, Ruhnka
and Young 7 hypothesize that an investor undertakes a twostep process.
First, the investments are screened according
to the potential magnitude of loss.
Those possessing an
acceptable level of risk in terms of the probability and
magnitude of potential loss are then evaluated in terms of
highest expected gain or maximum expected utility.
Thus,
downside risk becomes an important determinant of investment
behavior and an important subject for investment behavior
research.
7_
,
"Some Hypotheses about Risk in Venture Capital
Investing," J. of Business Venturing, 1991, v.6, pp.115-133.
36
4.3 The Importance of Downside Risk
The empirical studies
discussed above provide evidence that investors behavior
does not fit traditional expected utility theory.
The
traditional CAPM definition of risk as variability of
expected returns has been challenged by some researchers.
Sortino and Van der Meer 8 observe that investors are not
disappointed about variability of returns on the upside.
They argue that downside variability is more important in
decision-making than variability in general.
Other academicians have also explored the way downside
risk affects investment decision-making.
Arnott and
Bernstein9 believe that elimination of risk does not refer
to elimination of variability; rather, risk refers to the
risk of having insufficient assets to meet obligations or to
achieve desired minimum returns.
Hagigi and Klugerl0 fault the traditional risk definition
and offer a "safety first" rule which avoids downside risk.
Such models hold intuitive appeal, especially for real
8
"Downside Risk," J. of Portfolio Management, 1991,
,
v.17, n.4, pp.17-21.
9
"The Right Way to Manage Your Pension Fund,"
,
Harvard Business Review, 1988, n.1, pp.95-102.
10
,
"Safety First: An Alternative Performance
Measure," J. of Portfolio Management, 1987, v.13, n.4, pp.34-40.
37
estate portfolio managers who strive to beat a real estate
market proxy or invest to meet some minimum return.
A typical simplification of the traditional investment
models assumes that security returns are normally
distributed.
Most practitioners would agree that the
probability of investment returns is more often skewed, with
more variability on the upside as shown in Exhibit 4.2.
If
risk is defined as variability of return, then the best
possible outcome, +1,000%, would be deemed the most risky
outcome.
Exhibit 4.2: Probability Distribution of Expected Returns
-100%
+1,000%
0%
Expected Return
Sortino and Van der Meer introduce the concept of Minimal
Acceptable Return, which is described as the minimum return
required to accomplish some investment goals.
Only
potential returns which fall below the MAR are defined as
risky, and the farther below they fall, the greater the
risk.
Variation above the MAR and variation below the MAR
can be explicitly considered.
Other scenarios in a portfolio context underscore the
shortcomings of the traditional definition of risk.
In a
two asset portfolio, traditional definitions of risk
mitigation in an MPT context would strive to minimize
variability of returns.
Exhibit 4.3 shows how returns on
two assets are expected to vary over time.
Exhibit 4.3: Portfolio Returns
Asset X
(D
Portfolio
CT-
(D
(D
Asset Y
Time
In order to minimize variability of returns, an investor
should invest equal amounts in asset X and asset Y.
According to Markowitz' mean-variance efficiency (a
cornerstone of MPT),
the combined portfolio is preferable to
holding only asset X, because its expected volatility is
zero.
This is clearly misleading, because all investors
should choose to invest only in Asset X regardless of risk
tolerance or MAR.
Another illustration points out the problem of identifying
riskless assets in terms of variability.
As Exhibit 4.4
shows, combining assets A and B in a portfolio would provide
a portfolio with zero variance and low volatility about the
mean; however, the better investment is Asset D.
Clearly,
traditional measures of volatility do not capture the
distinct effect of downside volatility.
Corporate pension
fund advisors, subject to the Employee Retirement Income
Security Act (ERISA) of 1974, became particularly concerned
with meeting minimum investment goals by avoiding downside
risk.
Although ERISA's "prudent man" rule was initially
misinterpreted among pension fund trustees to imply
avoidance of downside volatility on an asset-by-asset basis,
subsequent clarifications by the Department of Labor relaxed
this strict interpretation.
Nevertheless, pension fund
sponsors remain concerned with meeting minimum return
objectives.
Exhibit 4.4: Volatility Around the Mean
(D
rr
.. ,Portfolio
-
Asset B
Time
A measure of the risk of falling below the MAR is provided
by Fishburnll.
His calculation of downside variance is
based on a probability-weighted function of deviations below
a specified level, the MAR. The model addresses the
weaknesses of the traditional mean-variance risk model while
remaining compatible with Modern Portfolio Theory (MPT).
Exhibit 4.5 illustrates the difference in downside risk
Although both assets have the same expected
measures.
return of 10% and the same standard deviation of 6%, the two
assets clearly are not equivalent.
Asset B has much more
downside variance.
11
,
"Mean-Risk Analysis with Risk Associated with
Below Target Returns," Am. Economic Review, 1977, v.67, n.2,
pp.116-125.
41
Exhibit 4.5: Comparison of Risk Measures
Asset A
Downside Variance
-8%
-16%
0%
8%1
16%
24%
A
Average Shortfall
Asset B
Downside Variance
-8%
-16%
0%
(-6.5%)
8%1
Minimal Acceptable Return
(10%)
16%
(-25.6%)
24%
Expected Return (12%)
Research by Harlow and Rao 12 demonstrate that the
downside variance measure can be incorporated into a CAPM
Whereas the traditional CAPM models variability
framework.
around the mean, the downside variance method models
downside variability below any arbitrary level.
Conceptually, the downside variance model is simply a more
general version of the traditional CAPM.
It allows
practitioners to measure risk as variability below an
arbitrary level, rather than measuring risk below the market
index.
12
,
"Asset Pricing in a Generalized Mean-Lower Partial
Moment Framework: Theory and Evidence," J. of Financial and
Quantitative Analysis, 1989, v.24, n.3, pp.285-310.
42
4.4 Application to Real Estate
The discussion of downside
risk, or volatility of returns below some minimal acceptable
return, has relevance to real estate decision-makers. The
assumption of normally distributed returns, and the
smoothing of classic expected utility analysis is avoided by
explicit consideration of downside risk.
The initial stages
of the risk management process, risk identification and
categorization process can benefit from explicit
consideration of downside risk by increasing the precision
of sensitivity analysis. Additionally, the downside variance
model is similar in form to the traditional CAPM, so
practitioners can integrate the model into the familiar
Modern Portfolio Theory milieu.
Chapter 5
A Consumption-Based Capital Asset Pricing Model
Chapter Summary: The consumption-based CAPM,
a generalized form of the CAPM which relates
an asset's price to its covariance of returns
to the level of national consumption, is
described. Its practical and theoretical
advantages over the traditional CAPM are
noted.
5.1 Introduction
The previous chapters of this thesis
explored some of the problems in attempting to apply
securities investment models to real estate.
The
dissimilarity of the asset classes and the lack of
comparable returns data for real estate remain major
stumbling blocks of this effort.
Consequently, real estate
risk identification and categorization techniques within
these frameworks are subject to the same conceptual and
practical limitations.
Recent research by David M. Geltner13 attempted to bridge
both the informational and conceptual gaps by modifying the
traditional investment models to incorporate observable data
13
,
"Estimating Real Estate's Systematic Risk from
Aggregate Level Appraisal-Based Returns," AREUEA Journal, 1989,
v.17, n.4, pp.463-481.
and aid investment decision-making for unsecuritized real
estate.
This suggests that the risk identification and
categorization process can use observable cash flow, lease
term, and market data, along with consistent indices
released by the U.S. government, to quantify risk.
As discussed in Chapter 1, the Capital Asset Pricing Model
(CAPM) has long been the favored investment model for
securities analysts.
Attempts to apply the CAPM to real
estate has necessitated the estimation of a "market return"
index for real estate.
Although the Russell-NCREIF Index
attempts to give a normalized view of investment returns,
there are serious data limitations to the practical
implementation of the CAPM to real estate.
Geltner suggests
that a more generalized version of the CAPM, the Consumption
CAPM (CCAPM) developed by Breeden14 , provides a more
pragmatic way to evaluate the risk characteristics of
unsecuritized real estate.
5.2 The Capital Asset Pricing Model
The CAPM is an
investment theory which describes the way prices for assets
are determined in a two-period time frame.
to minimize risk while maximizing return.
Investors seek
Assuming that
markets are efficient, then the "market" (a portfolio
14
,
"An Intertemporal Asset Pricing Model with
Stochastic Consumption and Investment Opportunities," J. of
Financial Economics, 1979, v.7, pp.265-296.
45
consisting of all risky assets within the market) will
represent an optimal risk/reward tradeoff.
Each asset's
measure of risk is captured by the statistic beta (B), which
estimates the asset's expected variability from the market
B is estimated by observing historical covariance
return.
to the market.
The following simplified equation summarizes
how this relationship is applied:
Exhibit 5.1: The Capital Asset Pricing Model
Re = Rf + B(R, - Rf)
Where:
Re = expected return
Rf = risk-free investment return
Rm = the market return
B = beta = the level of risk
For example, suppose the stock market's expected rate of
return is 10%, and the risk-free investment rate (the yield
on Treasury bills) is 6%.
If the stock of General Motors
has a 8 of 1.5 based on its historic performance relative to
the market, then the CAPM estimates General Motor's stock
return to be:
6% + 1.5(10% -
6%) = 12%.
5.3 The Consumption-Based CAPM
The CCAPM is a more general
form of the traditional CAPM. The CAPM uses an index of
stock market returns to derive an estimate of an individual
stock's risk.
Similarly, applying the CAPM to real estate
necessitates the estimation of a real estate market index of
returns.
46
The CAPM uses stock market returns as an index of wealth,
and the attempted application of CAPM to real estate
similarly has used real estate market returns as an index of
wealth.
This implies that total wealth is contained within
the asset class, whether it is stocks or real estate.
Although many investment management firms do separate
investment decisions by asset class, this fragmented view
misses the point of the CAPM.
Wealth is spread among many
asset classes, and the CAPM, too should be applied at the
portfolio level consisting of many asset classes.
The CCAPM, in contrast, generalizes that wealth beyond a
portfolio of stocks or a portfolio of real estate.
consumption serves as a proxy for total wealth.
National
The CCAPM
uses a readily available and objective data stream, national
consumption as reported quarterly by the U.S. Bureau of
Economic Analysis, to estimate individual wealth.
The concept behind the CCAPM is simple.
Utility is
defined as consumption, and less volatility of consumption
results in greater utility.
When national consumption is
greater than expected (as reported by the government
consumption index),
individuals' consumption is greater than
expected, and individuals are better off.
Similarly, when
the level of national consumption is below expectations,
individuals are worse off.
Any asset that achieves higher
47
returns when consumption is high, or achieves lower returns
when consumption is low increases individuals' consumption
volatility and results in lower individual utility.
Conversely, an asset that achieves higher returns when
consumption is below expectations and that achieves lower
returns when consumption is higher than expected will
decrease the volatility of individual consumption and
increase individual utility.
Thus, an asset that can smooth
an individual's expected consumption pattern will be worth
more than an asset which does not.
As mentioned in Chapter 1, recent research by Fama and
French has questioned the robustness of the CAPM as it
applies to the stocks.
Their arguments against beta, the
estimate of a stock's volatility relative the stock market
volatility, do not refute the concept of the CCAPM.
The
researchers argued that estimation of volatility of an asset
within its asset class (one stock's volatility within the
stock market) is not a strong indication of future
volatility.
The CCAPM avoids this problem by relating an
individual asset's volatility (a real estate investment) to
an aggregate measure of volatility of wealth (the national
consumption.)
5.4 Application to Real Estate Investment
The CCAPM
provides an opportunity to apply traditional investment
48
techniques to unsecuritized real estate.
The earlier
problem of finding an adequate real estate market index (as
a proxy for wealth) is circumvented by generalizing the CAPM
to use national consumption as a proxy for wealth.
The
historical property return stream, often known or reasonably
estimated, is also used. Thus, the data adequacy problem may
be alleviated using the CCAPM.
Ideally, application of the CCAPM to real estate would
resemble a simple equation similar to the CAPM equation of
Exhibit 5.1.: Re = Rf + 8(Rm - Rf).
Rm,
the return on the
market, would instead be the estimated change in national
consumption (using seasonally-adjusted quarterly personal
consumption expenditures from the Bureau of Economic
Analysis.)
B would be an estimate of the asset's return
variability relative to Rm,
estimated by measuring the
asset's ex post covariation with the changes in national
consumption.
Based on these estimates, one could estimate
the ex ante value of the real estate asset.
Of course, there remain a number of obstacles in the
practical application of the CCAPM.
First, although the
preliminary empirical study by Geltner showed that the CCAPM
may hold for valuing unsecuritized real estate, the size and
scope of the study was narrow.
Second, the model must be
specified more precisely; Geltner does not specify the
optimal number of time period lags between real estate
returns and market consumption.
Third, the real estate
return data, though more "knowable" than other data streams
required in other models, are not consistent throughout the
industry and require substantial preparation before use in
the model.
Thus, although much needs to be done, the CCAPM
appears a model worthy of continued research.
50
Chapter 6
Conclusions
This thesis attempted to describe tools outside of the
real estate mainstream which may be useful to a wide variety
of real estate decision-makers.
Emphasis was placed on
applying these tools to the early stages of the risk
management process: the risk identification and
categorization stages.
By focusing on the front-end, this
thesis attempted to introduce broad concepts to a field in
which both academics and practitioners express
dissatisfaction with the status quo techniques of risk
management.
The models described in Chapters 2 to 5 represent a wide
variety of conceptual bases, as different from each other as
from current analytical techniques.
Although each
represents a departure from the status quo, none is wholly
inconsistent with MPT. This is important to a field which is
rooted in the securities investment models.
The comparative advantage model of Chapter 2 provides a
broad framework which stresses the multi-party aspect of
51
real estate investment and development decisions.
The
methodology is qualitative and does not necessarily have the
new or better data requirements which have slowed
implementation of other models to real estate.
The extension of options pricing models to real estate
recognizes the option-like characteristics of real estate
and attempts to model them for decision-makers.
To apply
the technique, its quantitative approach demands a precision
in its input data which is difficult to achieve in real
estate.
However, this tool may provide helpful
approximations and estimates.
The modifications to expected utility theory and explicit
consideration of downside risk explored in Chapter 4 suffer
from the same limitations as its predecessor.
Expected
utility theory is a simple and well-understood investment
concept, but its practical application is limited by its
simplicity.
The consumption CAPM of Chapter 5 requires more rigorous
testing to determine its robustness.
The generalization of
the CAPM has intuitive appeal for its application to
unsecuritized real estate valuation.
Overall, the continuing search by practitioners and
academics for models of risk identification and
categorization underscores the difficulty of such an
undertaking.
The most significant aspect of this thesis,
then, may be in its approach: an attempt to borrow respected
concepts from other investment decision fields and apply
them in a meaningful way to real estate.
Bibliography
Apgar, M.,"A Strategic View of Real Estate," Real Estate
Issues, 1986, v.11, n.2, pp.6-11.
Arnott, R.D., and Bernstein, P.,"The Right Way to Manage
Your Pension Fund," Harvard Business Review, 1988, v.88,
n.1, pp.95-102.
Bawa, V.S., and Lindenberg, E.B.,"Capital Market Equilibrium
in a Mean-Lower Partial Moment Framework," J. of Financial
Economics, 1977, v.5, pp.189-200.
Blitzer, C.R., Lessard, D., and Paddock, J.L.,"Risk Bearing
and the Choice of Contract Forms for Oil Exploration and
Development," working paper, 1982, MIT International Energy
Studies Program.
Breeden, D.T., "An Intertemporal Asset Pricing Model with
Stochastic Consumption and Investment Opportunities," J. of
Financial Economics, 1979, v.7, pp.265-296.
Cole, R., Guilkey, D. and Miles, M., "Pension Fund
Investment Managers' Unit Values Deserve Confidence," Real
Estate Review, 1987, v.17, n.1, pp.84-89.
et. al. and Webb, B.,
"More Scientific Diversification
Strategies for Commercial Real Estate," Real Estate Review,
1989, v.19, n.1, pp.59-66.
Del Casino, J., "On Assembling Real Estate Portfolios," Real
Estate Issues, 1985, Fall/Winter, pp.47-49.
Fama, E., and French, K., "The Cross-Section of Expected
Stock Returns," J. of Finance, 1992, v.42, n.2, pp.427-465.
Firstenberg, P., Ross, S., and Zisler, R., "Real Estate: The
Whole Story," J. of Portfolio Management, 1988, v.14, n.3,
pp.22-34.
, and Wurtzebach, C.,
"Managing Portfolio Risk and
Reward," Real Estate Review, 1989, v.19, n.2, pp.61-65.
Fishburn, P., "Mean-Risk Analysis with Risk Associated with
Below-Target Returns," Am. Economic Review, 1977, v.67, n.2,
pp.116-125.
Geltner, D., "Estimating Real Estate's Systematic Risk from
Aggregate Level Appraisal-Based Returns," AREUEA Journal,
1989, v.17, n.4, pp.463-481.
"Risk and Returns in Commercial Real Estate: An
,
Exploration of Some Fundamental Relationships," 1989, Ph.D.
thesis, MIT.
Hagigi, M., and Kluger, B., "Safety First: An Alternative
Performance Measure," J. of Portfolio Management, 1987,
v.13, n.4, pp.34-40.
Harlow, W., and Rao, R., "Asset Pricing in a Generalized
Mean-Lower Partial Moment Framework: Theory and Evidence,"
J.of Financial and Quantitative Analysis, 1989, v.24, n.3,
pp.285-310.
Kahneman, D., and Tversky, A., "Prospect Theory: An Analysis
of Decision Under Risk," Econometrica, 1979, v.47, pp.263291.
Kloppenburg, W., "Real Estate Portfolio Management: One
Approach for Diversification of Specific Risk," 1990,
Master's thesis, MIT.
Lessard, D., "The Global Financial Revolution and the New
Development Finance," 1992, unpublished working paper, MIT.
, "Alternative Finance for Less Developed Countries: A
Primer," unpublished working paper, 1991, MIT.
,
"Evaluating Foreign Projects: An Adjusted Present
Value Approach," 1979, unpublished working paper, MIT.
Louargand, M., "Characteristics of Real Estate Investment
Risk," 1988, unpublished paper, MIT.
Lusht, K., "The Real Estate Pricing Puzzle," AREUEA Journal,
1988, v.16, n.2, pp.95-104.
Majd, S., and Pindyck, R., "Time to Build, Option Value, and
Investment Decisions," J. of Financial Economics, 1987,
v.18, pp.7-27.
Markowitz, H., Portfolio Selection - Efficient
Diversification of Investments, 1959, Yale University Press,
New Haven.
O'Connor, J., "Real Estate Development: Investment Risks and
Rewards," Real Estate Issues, 1986, Spring/Summer, pp.6-il.
Roulac, S., and Cirese, R., "A Risk Analysis Matrix to
Improve Investment Decisions," Real Estate Review, v.16,
pp.36-40.
Ruhnka, J., and Young, J., "Some Hypotheses about Risk in
Venture Capital Investing," J. of Business Venturing, 1991,
v.6, pp.115-133.
Pindyck, R., "Irreversibility, Uncertainty, and Investment,"
J. of Economic Literature, 1991, v.29, pp.1110-1148.
, "Irreversible Investment, Capacity Choice, and the
Value of the Firm," Am. Economic Review, 1988, v.78, n.5,
pp.969-985.
Pyhrr, S., and Cooper, J., Real Estate Investment: Strategy,
Analysis, Decisions, 1982, John Wiley & Sons, New York.
Shukla, R., and Trzcinka, C., "Research on Risk and Return:
Can Measures of Risk Explain Anything?" J. of Portfolio
Management, 1991, Spring, pp.15-21.
Sortino, F., and Van der Meer, R., "Downside Risk," J. of
Portfolio Management, 1991, v.17, n.4, pp.17-21.
Trigeorgis, L., "A Log-Transformed Binomial Numerical
Analysis Method for Valuing Complex Multi-Option
Investments," J. of Financial and Quantitative Analysis,
1991, v.26, n.3, pp.309-326.
Vernor, J., An Introduction to Risk Management in Property
Development, 1981, Urban Land Institute, Washington, D.C..
Webb, J., and Rubens, J., "Portfolio Considerations in the
Valuation of Real Estate," AREUEA Journal, 1986, v.14, n.3,
pp.465-495.
Zerbst, R., and Cambon, B., "Real Estate: Historical Returns
and Risks," J. of Portfolio Management, 1984 Spring, pp.520.
56